Feature Selection for High Dimensional and Complex Data

Variable and feature selection has been a research topic with practical significance in many areas such as statistics, pattern recognition, machine learning, and data mining. During the past years, more and more high-dimensional datasets with small number of samples and complex data (e.g. graphs, strings, documents,..) are emerging, which have posed unprecedented challenges to data mining and pattern recognition. In order to construct simpler and more comprehensible models, improve data mining performance, and help to prepare, clean, and understand these high dimensional and complex data, the existed feature selection methodologies need to be adapted or new methodologies should be developed.

The session aims to further the cross-discipline, collaborative effort in feature selection research and share the methods. The topics include, but not limited to: Feature ranking, Feature extraction/construction, Selection in small samples with high dimensionality, Selection in extremely high-dimensional domains, Kernel feature selection, Semi-supervised feature selection, Ensemble feature selection, Combination of feature selection and feature extraction, Novel evaluation criterion for feature selection, Real-world case studies and application, such as gene selection, web mining, text processing, bioinformatics, social networks, etc.

Neural information processing and neural coding

This special session proposal in ISNN2010 addresses problems related to the information representation and processing in the biological nervous system, which is fundamental and important for develop atificial neural networks and AI. It brings together researchers working across this field to share their experience and discuss advanced thinking. The introduction and discussion of neural information processing and neural coding will conduce to development of artificial neural networks and AI. The special session will be of great value to ISNN2010 participants.

Computational Intelligence for Robot Brain

The theme reflects the growing development of robot intelligence in all aspects of human behavior and their increasing diversity. This session is motivated by that view and is centered around the study of various aspects of human actions since these are intimately linked with many higher cognitive abilities.

Natural language processing and machine learning

Natural language processing (NLP) is a key technology in the information processing area and has a wide range of applications from word processor to information access including information retrieval, machine translation, text categorization, text mining, and so forth. Studies on NLP based on neural networks as a machine learning method began in the early 1980s with studies on implementing semantic networks, word-sense disambiguation, anaphora resolution, and parsing. Since then, with the increase in NLP research based on very large corpora, machine learning has attracted a great deal more attention from both the NLP and machine learning researchers. The special session provides a forum for researchers interested in further advancing the state of the art in developing NLP techniques that use machine learning.